| dc.description.abstract | In the pharmaceutical industry, pooled human liver microsomes (HLM) and pooled cryopreserved human hepatocytes (CHH) are the most commonly used test systems to measure the in vitro metabolic intrinsic clearance (CLint) of investigational new drugs in order to identify drug candidates with favorable pharmacokinetic properties such as once-a-day dosing and low oral dose. However, both HLM and CHH have been shown to underpredict the in vivo clearance of drugs. For drugs whose clearance is predominantly determined by P450 enzymes, metabolic clearance in CHH would be expected to equal that in HLM even if both test systems underpredicted (or overpredicted) in vivo clearance. Curiously, in the case of drugs with high intrinsic clearance (CLint), CHH underpredict in vivo clearance to a greater extent than HLM. Such system-dependent clearance has been reported for midazolam, a high CLint drug that is the most widely used CYP3A4/5 substrate for both the in vitro and in vivo assessment of drug-drug interactions (DDIs). Previous investigators have proposed that the system-dependent clearance of midazolam is due to permeability- or cofactor-restricted clearance in CHH (i.e., clearance in hepatocytes is limited by membrane permeability or the availability of NADPH). The objective of this dissertation research was to determine the mechanism underlying the system-dependent clearance of midazolam. Studies of midazolam clearance in HLM and CHH confirmed previous reports that midazolam clearance is almost an order of magnitude lower in CHH than HLM, a system-dependent difference that was much more pronounced than with other CYP3A4/5 substrates (namely alfentanil, nifedipine and verapamil). In vitro to in vivo extrapolation (IVIVE) of clearance established that HLM accurately predicted the in vivo clearance of midazolam whereas CHH underpredicted midazolam clearance by a factor of 5. Permeabilizing CHH by sonication or treatment with the pore-forming agent saponin did not increase the rate of midazolam metabolism, even in the presence of excess NADPH. Furthermore, the rate of midazolam uptake by CHH was found to greatly exceed the rate of midazolam metabolism, and microsomes isolated from pooled CHH had comparable CYP3A4/5 activity towards midazolam as microsomes prepared directly from human liver. These results suggested that neither membrane permeability nor intracellular cofactor availability were likely explanations for the system-dependent clearance of midazolam. The impact of in vitro incubation conditions on P450 activity, namely the ionic strength of the incubation buffer and the effect of cell culture media, was evaluated as a possible explanation for the system-dependent clearance of midazolam. As part of this investigation, a cell culture medium was sought that was capable of increasing midazolam clearance in CHH. Compared with KHB (the medium used in the initial experiment), Williams’ E medium supported similar rates of midazolam metabolism but the other three media examined, namely Waymouth’s, MCM+ and DMEM, supported lower rates of midazolam metabolism. In other words, none of the media examined corrected the system-dependent clearance of midazolam (and three of them made matters worse). In general, P450 activities in HLM were maximal at 50 mM phosphate buffer, with the exception of CYP3A4/5 and CYP2E1, where the enzymatic activities increased with increasing buffer ionic strength. The activity of these two enzymes was markedly decreased when HLM were incubated in Waymouth’s, MCM+ or DMEM (the same three media that decreased the rate of midazolam metabolism in CHH). The effect of certain cell culture media on reducing midazolam clearance in both HLM and CHH was not observed with other CYP3A4/5 substrates (namely alfentanil, nifedipine, verapamil, testosterone and atorvastatin). The kinetics of midazolam metabolism in HLM in the presence of various cell culture media suggested that both MCM+ and Williams’ E medium contained an inhibitory substance as evidenced by a marked increase in Km compared with 50 mM phosphate buffer. Studies with complete versions or salt-only versions of each medium on CYP3A4/5 activity in HLM further suggested the presence of an inhibitory component in certain cell culture media. Although, this dissertation research was unsuccessful in identifying a the cell culture medium that corrected the system-dependent clearance of midazolam, it disproved two previous explanations for this phenomenon and formed the basis for recommending valuable improvements for the conduct of in vitro metabolism studies, such as ideal buffering conditions for HLM and the use of Williams’ E media for studies in CHH. In an effort to identify a CYP3A4/5 substrate with the same system-dependent clearance characteristics as midazolam, the CYP3A4/5 substrate loratadine was examined in CHH. The rate of conversion of loratadine to desloratadine by CHH varied depending on the culture medium in a manner similar to that observed with all other CYP3A4/5 substrates except midazolam. In these studies of loratadine metabolism in hepatocytes, 3-hydroxydesloratadine was detected, which was unexpected because no prior in vitro test system (such as HLM or recombinant CYP enzymes) or non-clinical species in vivo had been previously shown to support its formation. 3-Hydroxydesloratadine is the major human metabolite of desloratadine, and the enzyme responsible for its formation had not been identified despite this being a postmarketing requirement imposed by the FDA on the manufacturer Schering-Plough. Capitalizing on the detection of 3-hydroxydesloratadine in incubations with CHH, studies were conducted to elucidate the enzymology of 3-hydroxydesloratadine formation and to evaluate the potential of desloratadine to be the victim or perpetrator of drug interactions. These studies established that the conversion of desloratadine to 3-hydroxydesloratadine is catalyzed by CYP2C8 but only after desloratadine is converted to an N-glucuronide by UGT2B10. Formation of 3-hydroxydesloratadine could be blocked with inhibitors of either CYP2C8 or UGT2B10. Recombinant CYP2C8 formed 3-hydroxydesloratadine only when co-incubated with recombinant UGT2B10 and HLM formed 3-hydroxydesloratadine only when supplemented with both NADPH and UDP-GlcUA (the cofactors to support CYP and UGT enzymes). The formation of 3-hydroxydesloratadine was proposed to follow a three step process: N-glucuronidation of desloratadine by UGT2B10, followed by 3-hydroxylation by CYP2C8 and a de-conjugation event to form 3-hydroxydesloratadine. Desloratadine was found to be a weak inhibitor of CYP2B6, CYP2D6 and CYP3A4/5, but a potent and selective inhibitor of UGT2B10. Further studies with inhibitors of UGT2B10 and UGT1A4 established that UGT2B10 is the sole UGT responsible for supporting the CYP2C8-dependent formation of 3-hydroxydesloratadine. In addition to solving a long-standing mystery surrounding the enzymology of 3-hydroxydesloratadine formation, the results of this dissertation have implications on the labeling of desloratadine and, more importantly, provide a pathway for investigating the genetic basis of the impaired metabolism of desloratadine (the so-called poor metabolizer phenotype) observed in a small percentage of patients taking desloratadine. Furthermore, the identification of desloratadine as a UGT2B10 selective inhibitor advances the field of UGT research towards the goal of identifying a selective inhibitor of each UGT enzyme for use in in vitro metabolism studies. | |
